Advertisement
Research ArticleArticle

In Vitro Evaluation of the Reductase Activities of Human AKR1C3 Allelic Variants

Shiori Takano, Tatsuki Fukami, Hiroyuki Ichida, Kohei Suzuki, Masataka Nakano and Miki Nakajima
Drug Metabolism and Disposition September 2023, 51 (9) 1188-1195; DOI: https://doi.org/10.1124/dmd.123.001264

Abstract

Aldo-keto reductase 1C3 (AKR1C3) plays a role in the detoxification and activation of clinical drugs by catalyzing reduction reactions. There are approximately 400 single-nucleotide polymorphisms (SNPs) in the AKR1C3 gene, but their impact on the enzyme activity is still unclear. This study aimed to clarify the effects of SNPs of AKR1C3 with more than 0.5% global minor allele frequency on the reductase activities for its typical substrates. Recombinant AKR1C3 wild-type and R66Q, E77G, C145Y, P180S, or R258C variants were constructed using insect Sf21 cells, and reductase activities for acetohexamide, doxorubicin, and loxoprofen by recombinant AKR1C3s were measured by liquid chromatography–tandem mass spectrometry. Among the variants tested, the C145Y variant showed remarkably low (6%–14% of wild type) intrinsic clearances of reductase activities for all three drugs. Reductase activities of these three drugs were measured using 34 individual Japanese liver cytosols, revealing that heterozygotes of the SNP g.55101G>A tended to show lower reductase activities for three drugs than homozygotes of the wild type. Furthermore, genotyping of the SNP g.55101G>A causing C145Y in 96 Caucasians, 166 African Americans, 192 Koreans, and 183 Japanese individuals was performed by polymerase chain reaction–restriction fragment length polymorphism. This allelic variant was specifically detected in Asians, with allele frequencies of 6.8% and 3.6% in Koreans and Japanese, respectively. To conclude, an AKR1C3 allele with the SNP g.55101G>A causing C145Y would be one of the causal factors for interindividual variabilities in the efficacy and toxicity of drugs reduced by AKR1C3.

SIGNIFICANCE STATEMENT This is the first study to clarify that the AKR1C3 allele with the SNP g.55101G>A causing C145Y results in a decrease in reductase activity. Since the allele was specifically observed in Asians, the allele would be a factor causing an interindividual variability in sensitivity of drug efficacy or toxicity of drugs reduced by AKR1C3 in Asians.

Introduction

Compounds containing ketone, aldehyde, or quinone groups are subjected to a reduction reaction by members of the aldo-keto reductase (AKR) and short-chain dehydrogenase/reductase (SDR) families (Barski et al., 2008; Malátková et al., 2010). The human AKR family consists of 15 isoforms (https://akrsuperfamily.org/). Among them, 12 isoforms, AKR1A1, 1B1, 1B10, 1B15, 1C1, 1C2, 1C3, 1C4, 1D1, 1E2, 7A2, and 7A3, have been reported to be involved in the reduction reaction of xenobiotics (Fukami et al., 2022). The human SDR family consists of 75 isoforms (Persson and Kallberg, 2013), and seven isoforms, carbonyl reductases (CBRs) 1, 3, 4, 11β-hydroxysteroid dehydrogenase type 1 (HSD11B1), 17β-hydroxysteroid dehydrogenase type 12 (HSD17B12), dehydrogenase/reductase member 4, and dicarbonyl/L-xylulose reductase, have been suggested to be involved in the reduction reaction of xenobiotics (Matsunaga et al., 2006, Ichida et al., 2023a). Among these members, AKR1C1, AKR1C2, AKR1C3, AKR1C4, CBR1, and HSD11B1 are highly expressed in the liver and contribute to the reduction reaction of clinically used drugs (Barski et al., 2008; Amai et al., 2020) with overlapping substrate specificities (Ichida et al., 2023a). For example, the antipsychotic drug haloperidol is reduced by CBR1, and the nonsteroidal anti-inflammatory drug loxoprofen is reduced by AKR1C3, AKR1C4, and CBR1, with varied contribution rates in the human liver (Ichida et al., 2023a). In addition, we recently found that the nonsteroidal anti-inflammatory drug nabumetone is reduced by HSD17B12, which is also highly expressed in the liver and was not previously recognized as a drug-metabolizing reductase (Ichida et al., 2023b).

AKR1C3 is also known as 17β-HSD type 5 (HSD17B5) or prostaglandin F2α synthase since it catalyzes the conversion of Δ4-androstene-3,17-dione to testosterone, 5α-androstane-3,17-dione to 5α-dihydrotestosterone, and estrone to 17β-estradiol, as well as the conversion of PGH2 to PGF2α and PGD2 to 11β-PGF2α (Penning, 2015). In drug metabolism, AKR1C3 catalyzes the reduction reaction of several drugs, such as warfarin and ketodarolutamide (Malátková et al., 2016, Taavitsainen et al., 2021). In addition to these two compounds, we recently found that AKR1C3 participates in the reduction reaction of acetohexamide, doxorubicin, and loxoprofen in the human liver, with higher contributions (>50%) than previously expected (Ichida et al., 2023a), although these reduction reactions had been considered to be catalyzed predominantly by CBR1 (Ohara et al., 1995, Kassner et al., 2008). Thus, the significant role of AKR1C3 in drug metabolism has been progressively clarified.

Interindividual differences in drug efficacy and adverse reactions are caused by the variability in drug metabolism due to a variety of factors, including differences in gene expression and enzyme activity, drug-drug interactions, and genetic polymorphisms. As for AKR1C3, the gene expression is known to be regulated by several transcription factors, such as aryl hydrocarbon receptor (Yamashita et al., 2019), NF-E2–related factor 2 (Jung et al., 2013), Sp1/3 (Yamashita et al., 2019), and signal transducer and activator of transcription 3 (Zhou et al., 2020). In vitro studies showed that some of the nonsteroidal anti-inflammatory drugs such as naproxen and flufenamic acid strongly inhibit AKR1C3 (Byrns et al., 2008; Adeniji et al., 2012). Thus, information on gene regulation and enzyme inhibition for AKR1C3 has been accumulated. Since genetic polymorphism is also one of the factors responsible for the large interindividual differences, the effects of single-nucleotide polymorphisms (SNPs) of various types of drug-metabolizing enzymes on their activity have been studied with great enthusiasm (Evans and Mcleod, 2003). AKR1C3 is a polymorphic enzyme with more than 25,000 SNPs registered in the Single Nucleotide Polymorphism Database (dbSNP), and approximately 400 SNPs are located in the coding region (https://www.ncbi.nlm.nih.gov/snp/?term=akr1c3). Among them, there were 5 SNPs in coding regions with more than 0.5% global minor allele frequency as shown in Table 1. Several research groups (Bains et al., 2010, Platt et al., 2016, Detlefsen et al., 2022) have investigated the effects of SNPs other than g.55101G>A (C145Y) on reductase activities for different substrates by constructing recombinant enzymes. Their results were inconsistent between studies (Table 1), probably due to the differences in substrates, buffer composition, or pH in reactions. Therefore, we aimed to gain further insight into the effects of genetic variants, including the undetermined variant with g.55101G>A (C145Y), on the reductase activities of three clinical drugs with different structural properties: acetohexamide, doxorubicin, and loxoprofen.

View this table:
TABLE 1

Information about allelic variants of human AKR1C3 with more than 0.5% of gMAF registered on dbSNP database

gMAF values were from dbSNP database.

Materials and Methods

Chemicals and Reagents

Acetohexamide, doxorubicin hydrochloride, and loxoprofen sodium were purchased from FUJIFILM Wako Pure Chemical (Osaka, Japan). Hydroxyhexamide and doxorubicinol were purchased from MedChemExpress (Princeton, NJ) and Toronto Research Chemicals (Toronto, Canada), respectively. trans-Hydroxyloxoprofen was kindly provided by Daiichi Sankyo (Tokyo, Japan). Glucose-6-phosphate (G6P), β-nicotinamide adenine dinucleotide phosphate, and glucose-6-phosphate dehydrogenase were purchased from Oriental Yeast (Tokyo, Japan). PrimeSTAR Max Premix, Tks Gflex DNA polymerase, In-fusion HD Cloning Kit, and Afa I were purchased from Takara Bio (Shiga, Japan). OneTaq DNA polymerase was purchased from New England Biolabs (Bevebly, MA). The NucleoSpin Gel and PCR Clean-up Kit was purchased from MACHEREY-NAGEL (Düren, Germany). The Bac-to-Bac Baculovirus Expression System Kit was purchased from Invitrogen (Carlsbad, CA). IRDye 680 goat anti-mouse IgG antibody was purchased from LI-COR Bioscience (Lincoln, NE). The pFastBac1 AKR1C3 plasmid and recombinant AKR1C3 wild type expressed in baculovirus-infected insect cells were prepared previously (Ichida et al., 2023a). All other chemicals and solvents were of the highest quality commercially available.

Genomic DNA and Human Liver Cytosols

Genomic DNA was previously extracted from blood samples of 96 Caucasians, 166 African Americans, 192 Koreans, and 183 Japanese subjects (Hirosawa et al., 2021). The use of genomic DNA samples was approved by the Human Studies Committee of Washington University School of Medicine (St. Louis, MO), Soonchunhyang University Hospital (Chonan, Korea), and the Ethics Committee of Kanazawa University (Kanazawa, Japan) (certification number: 375; approval date: April 20, 2015). Human liver samples from 34 Japanese individual donors were obtained from Dokkyo Medical University (Mibu, Japan) and Iwate Medical University (Yahaba, Japan). The use of the human liver was approved by the Ethics Committee of Dokkyo Medical University and Iwate Medical University (certification number: 170; approval date: June 24, 2008). Donor information is listed in Supplemental Table 1. Cytosols of individual livers were prepared according to the method described previously (Kobayashi et al., 2012).

Construction of Recombinant Human AKR1C3 Variants Expressed in Sf21 Cells

Expression plasmids for AKR1C3 containing each SNP [c.197G>A (R66Q), c.230A>G (E77G), c.434G>A (C145Y), c.538C>T (P180S), or c.772C>T (R258C)] were constructed by site-directed mutagenesis with the pFastBac1 vector containing AKR1C3 wild-type cDNA (Ichida et al., 2023a) according to a previous study (Hirosawa et al., 2021) with slight modifications. To introduce the mutations, polymerase chain reaction (PCR) was performed using PrimeSTAR MAX Premix or Tks Gflex DNA polymerase with the primers shown in Table 2. The PCR products were purified using a NucleoSpin Gel and PCR Clean-up Kit and ligated using an In-fusion HD Cloning Kit. Nucleotide sequences were confirmed by DNA sequence analysis (AZENTA, South Plainfield, NJ). The pFastBac1 vectors with each AKR1C3 allelic variant were transformed into DH10Bac competent cells, and the insert was then transposed into bacmid DNA. A Bac-to-Bac Baculovirus Expression System was used to express human AKR1C3 variants in Spodoptera frugiperda (Sf21) cells. Finally, cell homogenates were prepared as described previously (Fukami et al., 2010) by suspension in buffer containing 10 mM Tris-HCl (pH 7.4), 1 mM EDTA (pH 7.4), and 20% glycerol. The Bradford method using γ-globulin as the standard was used to determine the protein concentrations of recombinant AKR1C3 variants (Bradford, 1976).

View this table:
TABLE 2

Primers used to introduce mutations in the AKR1C3 gene and for AKR1C3 genotyping

The underlined letters indicate the allelic substitutions.

Immunoblot Analysis

SDS-PAGE and immunoblot analysis for human AKR1C3 were in accordance with Laemmli (1970). Five micrograms of recombinant AKR1C3s or 30 μg of human liver cytosol (HLC) from 34 individual liver samples were separated on 10% polyacrylamide gels and electrotransferred to polyvinylidene difluoride membranes (Immobilon-P; Millipore Corporation, Billerica, MA). The membrane was probed with a mouse anti-human AKR1C3 monoclonal antibody (Sigma Aldrich, St. Louis, MO) and the corresponding fluorescent dye-conjugated secondary antibody. Band intensities were quantified using an Odyssey Infrared Imaging system (LI-COR Biosciences, Lincoln, NE). This analysis was performed within a linear range of band intensity versus protein amount. To confirm that the applied amount of protein did not vary between samples, Coomassie brilliant blue staining or immunoblotting with a polyclonal rabbit anti-human glyceraldehyde-3-phosphate antibody (Novus Biologicals, Centennial, CO) was performed when AKR1C3 levels in recombinant AKR1C3s or individual HLC samples were analyzed (Supplemental Figs. 1 and 2).

Reductase Activities for Acetohexamide, Loxoprofen, and Doxorubicin

Reductase activities of acetohexamide, loxoprofen, and doxorubicin were measured as follows: a typical incubation mixture (final volume of 0.2 mL) containing 100 mM potassium phosphate buffer (pH 7.4), enzyme sources (recombinant reductases or HLC), and the substrate was prepared. After preincubation at 37°C for 2 minutes, reactions were initiated by the addition of an NADPH-generating system (5 mM G6P, 0.5 mM β-nicotinamide adenine dinucleotide phosphate, 5 mM MgCl2, and 1 U/mL G6P dehydrogenase). After incubation at 37°C, the reaction was terminated by the addition of 0.1 mL of ice-cold acetonitrile. The mixture was then centrifuged at 20,380g for 5 minutes. For the measurement of acetohexamide, loxoprofen, and doxorubicin reductase activities, an LCMS-8040 (Shimadzu, Kyoto, Japan) equipped with an LC-20AD HPLC system was used according to the method previously reported by Ichida et al. (2023a), summarized in Supplemental Table 2. When recombinant AKR1C3 wild type or variants were used as the enzyme source, the protein concentrations were set at 0.1 mg/mL for acetohexamide and doxorubicin reductions and 0.2 mg/mL for loxoprofen reduction. For kinetic analyses, the concentrations of acetohexamide, loxoprofen, and doxorubicin were set at 0.5–100 µM, 1–500 µM, and 5–400 µM, respectively. The activity of nontransfected Sf21 cells was observed with 1%–25% of AKR1C3 wild-type activities (Supplemental Fig. 3). This would be due to endogenous reductases expressed in Sf21 cells because mock human embryonic kidney 293 cells did not show any acetohexamide reductase activity (Supplemental Fig. 4). Accordingly, the activity of each recombinant AKR1C3 was calculated by subtracting the activity by nontransfected Sf21 cells. When individual HLC samples were used as the enzyme source, protein concentrations were set at 0.2 mg/mL in all reactions. The concentrations of acetohexamide, loxoprofen, and doxorubicin were set at 20, 5, and 15 µM, respectively, which were the predicted hepatic maximum unbound drug concentrations (Ichida et al., 2023a). For kinetic analysis, the concentration of acetohexamide was set at 1.5–100 µM. Kinetic parameters were determined by nonlinear regression analysis using GraphPad Prism version 5 (GraphPad Software, La Jolla, CA) or KaleidaGraph (Synergy Software, Reading, PA), with the following equations for substrate inhibition kinetics (eq. 1) or Michaelis-Menten kinetics (eq. 2) or combined equation of eq. 1 and eq. 2: Embedded Image Embedded Image where V is the velocity of the reaction, S is the substrate concentration, Km is the Michaelis–Menten constant, Vmax is the maximum velocity, and Ki is the substrate inhibition constant.

Genotyping of the Allele with g.55101G>A (rs28943579)

Genotyping of the allele with g.55101G>A, which causes an amino acid substitution C145Y in AKR1C3, was performed by PCR-restriction fragment length polymorphism analysis using genomic DNA samples from 96 Caucasians, 166 African Americans, 192 Koreans, and 183 Japanese subjects. PCR was performed using OneTaq DNA polymerase with the primers listed in Table 2. The PCR products were digested with Afa I. In electrophoresis, the wild type yielded 130- and 222-bp fragments, and the variant yielded a 352-bp fragment.

Statistical Analysis

Statistical analysis of the kinetic parameters was performed by two-tailed Student’s t test. Statistical analysis of reductase activities in HLC between different AKR1C3 genotypes was performed by the Mann-Whitney U test. P < 0.05 was considered statistically significant.

Results

Construction of Recombinant AKR1C3 Variants

Recombinant AKR1C3 wild type and variants with R66Q, E77G, C145Y, P180S, or R258C were constructed. By immunoblot analysis, the relative expression levels of AKR1C3 variants to wild type were as follows: R66Q, 1.22; E77G, 1.65; C145Y, 2.07; P180S, 1.14; and R258C, 2.04 (Fig. 1A). In a subsequent study, the activities of the recombinant AKR1C3 variants were normalized to these values.

Fig. 1.
Fig. 1.

Kinetic analyses of acetohexamide, loxoprofen, and doxorubicin reductase activities by recombinant AKR1C3 variants. (A) Expression of AKR1C3 protein in wild-type (WT) and five recombinant AKR1C3 variants analyzed by immunoblotting. Relative AKR1C3 protein expression levels to WT are shown above each band. Immunoblot analysis was performed in triplicate. Kinetic analyses of (B) acetohexamide, (C) loxoprofen, and (D) doxorubicin reductase activities by recombinant AKR1C3 variants. The enzyme activities of the recombinant AKR1C3 variants were normalized to the relative AKR1C3 protein expression level. The enzyme activities of recombinant AKR1C3 WT, R66Q, E77G, C145Y, P180S, and R258C are shown in black, orange, blue, red, green, and purple, respectively. Experiments were replicated in three independent incubations with a single batch of recombinant AKR1C3 variants. Each point represents the mean ± S.D.

Kinetic Analyses of Acetohexamide, Loxoprofen, and Doxorubicin Reductase Activities Using Recombinant AKR1C3

Using the constructed recombinant AKR1C3s, the reductase activities of acetohexamide, loxoprofen, and doxorubicin were measured. Among 5 AKR1C3 variants, C145Y showed a remarkably low intrinsic clearance (CLint) value for the reduction reaction of all three substrates. For acetohexamide reduction (Fig. 1B), the Km and Vmax values of the C145Y variant were 40.2- and threefold (P < 0.001) higher than those of the wild type, resulting in a 93% (P < 0.001) lower CLint value than that of the wild type. For loxoprofen reduction (Fig. 1C), the Km value of C145Y was 13.3-fold (P < 0.001) higher than that of the wild type, whereas the Vmax value was 21% (P < 0.01) lower than that of the wild type, resulting in a 94% (P < 0.001) lower CLint value than that of the wild-type. The Km and Vmax values for doxorubicin reductase activity of the C145Y variant were 2.2-fold (P < 0.001) higher and 70% (P < 0.001) lower, respectively, than those of the wild type (Fig. 1D), resulting in an 86% (P < 0.001) lower CLint value than that of the wild type. Interestingly, the activities of the AKR1C3 C145Y variant followed the Michaelis-Menten equation, whereas the AKR1C3 wild-type and other variants followed the substrate inhibition equation for acetohexamide and loxoprofen reductase activities (Fig. 1, B and C). However, for doxorubicin reduction (Fig. 1D), the activities of AKR1C3 wild type and all variants followed the Michaelis-Menten equation. The other AKR1C3 variants showed moderate or no effect on drug reductase activities. As an exception, the P180S variant showed 30% (P < 0.05) higher and 20% (P < 0.05) lower CLint values than the wild type for acetohexamide and doxorubicin reductase activities, respectively. The R258C variant showed a 1.5-fold (P < 0.01) higher CLint value than the wild type in doxorubicin reductase activity, with a 1.6-fold (P < 0.001) higher Vmax value (Fig. 1D).

Taken together, we found that AKR1C3 C145Y has substantially decreased enzyme activity regardless of the substrate.

Allele Frequencies of the Allele with g.55101G>A (C145Y) in 96 Caucasians, 166 African Americans, 192 Koreans, and 152 Japanese

The allele frequencies of the allele with g.55101G>A causing the C145Y substitution were examined in four ethnic groups. This allele was not found in Caucasians and African Americans (Table 4). In 192 Korean and 152 Japanese subjects, 26 and 13 subjects were heterozygotes of the allele, resulting in allele frequencies of 6.8% and 3.6%, respectively (Table 4). Thus, this allele was found specifically in Asians.

Comparison of Reductase Activities for Acetohexamide, Loxoprofen, and Doxorubicin between Carriers and Noncarriers of the Allele with g.55101G>A (C145Y)

Reductase activities for acetohexamide, loxoprofen, and doxorubicin were measured using 34 HLC samples (Japanese), including three carriers and 31 noncarriers of the allele with g.55101G>A. At 20 µM acetohexamide, the activities in HLC from heterozygotes of the mutant allele (12.0 ± 2.6 pmol/min per mg protein) were lower than those from noncarriers (16.9 ± 6.3 pmol/min per mg protein), although the difference was not statistically significant (Fig. 2A). Similar trends were observed for loxoprofen reduction (13.3 ± 3.9 versus 15.7 ± 5.6 pmol/min per mg protein at 5 µM substrate concentration) and doxorubicin reduction (2.2 ± 0.3 versus 2.7 ± 0.9 pmol/min per mg protein at 15 µM substrate concentration) (Fig. 2, B and C). AKR1C3 protein levels were not significantly different between noncarriers and carriers (Fig. 2D). Thus, the results imply that the carriers of this allele have decreased enzyme activity.

Fig. 2.
Fig. 2.

Difference in the reductase activities for acetohexamide, loxoprofen, and doxorubicin between HLC samples from subjects genotyped as noncarrier and heterozygote of g.55101G>A (rs28943579) allele. (A) Acetohexamide, (B) loxoprofen, and (C) doxorubicin reductase activities in 34 individual Japanese liver cytosols. Among 34 liver samples, three samples were derived from heterozygotes of g.55101G>A, and the remaining 31 samples were defined as homozygotes of AKR1C3 wild type. In each figure, substrate concentrations are set at (A) 20, (B) 5, and (C) 15 µM, which is the predicted hepatic maximum unbound drug concentration of each substrate. Statistical analysis was performed by the Mann‒Whitney U test. (D) AKR1C3 protein levels in 34 individual Japanese liver cytosols determined by immunoblotting. The relative band intensity of each sample was calculated by setting the band intensity of sample #25 as 1.00, which had the lowest value. Framed sample numbers represent heterozygotes of g.55101G>A, and the rest of the samples are noncarriers of g.55101G>A.

Kinetics of Acetohexamide Reductase Activity by HLC Sample from a Heterozygote of the Allele with g.55101G>A

Our recent study revealed that acetohexamide reduction in HLC is also catalyzed by CBR1 (Ichida et al., 2023a), but the contribution of AKR1C3 to acetohexamide reduction in HLC (78%) is higher than that of CBR1 (17%) at the predicted hepatic maximum unbound concentration (20 µM) based on the kinetics of AKR1C3 showing substrate inhibition with a Km value of 3.5 ± 0.2 µM (Fig. 1B) and the kinetics of CBR1 showing continuously increasing activities up to 100 µM substrate concentration (Ichida et al., 2023a). We investigated whether the kinetics of acetohexamide reductase activity by HLC from a donor with AKR1C3 g.55101G>A are different from that of a donor homozygous for AKR1C3 wild type. Samples #10 and #30 in Fig. 2A were selected because these samples had average acetohexamide reductase activities in each group. Eadie-Hofstee plots of acetohexamide reductase activities of #10 and #30 samples followed biphasic kinetics (Fig. 3A), suggesting the involvement of two or more enzymes in the reduction reaction of acetohexamide. The substrate concentration-velocity (S-V) plots of acetohexamide reductase activities of both samples followed the combined equation of substrate inhibition and Michaelis-Menten equations (Fig. 3B; Table 5), supporting the results of the kinetic profiles of recombinant AKR1C3 and CBR1 (Ichida et al., 2023a). The Vmax1 value obtained from the substrate inhibition equation in the #30 sample (9.2 ± 1.5 pmol/min per mg protein) was 45% (P < 0.05) lower than that in the #10 sample (16.6 ± 1.9 pmol/min per mg protein), although the Km1 values were not significantly different between #30 and #10 (18.0 ± 3.9 µM and 9.9 ± 1.7 µM, respectively) (Fig. 3B; Table 5), indicating that a donor of AKR1C3 g.55101G>A had lower activity than a donor of wild-type AKR1C3 homozygote in the low acetohexamide concentration range. The Km2 and Vmax2 values obtained from the Michaelis-Menten equation were similar between the #10 and #30 samples (Fig. 3B; Table 5). The result indicates that the activity component owing to CBR1 was similar between the two samples, although the kinetic parameters of CBR1 are inaccurate because the activity was not saturated at the substrate concentrations due to the limited solubility of acetohexamide. The CLint calculated from an initial slope of the velocity versus the substrate concentration in the #30 sample (0.97 ± 0.07 µL/min per mg protein) was 42% (P < 0.001) lower than that in the #10 sample (1.66 ± 0.05 µL/min per mg protein) (Table 5).

Fig. 3.
Fig. 3.

Comparison of (A) Eadie-Hofstee plots and (B) substrate concentration-velocity (S-V) plots of acetohexamide reductase activities by individual liver cytosols from subjects genotyped as noncarrier (sample #10) and heterozygote (sample #30) of g.55101G>A in Fig. 2. The enzyme activities of liver cytosols from #10 and #30 are shown in gray circles and black triangles, respectively. Each point represents the mean ± S.D. of triplicate determinations. (C) Simulated kinetics of acetohexamide reductase activities of the AKR1C3 component and CBR1 component in samples #10 and #30. Kinetic curves were drawn by putting each kinetic parameter into the substrate inhibition equation (eq. 1) or Michaelis-Menten equation (eq. 2). The simulated curves for enzyme activities of the AKR1C3 component and CBR1 component are shown with black and gray lines, respectively.

S-V plots of the acetohexamide reductase activities by each component for AKR1C3 or CBR1 were simulated by putting the kinetic parameters shown in Table 5 into the substrate inhibition equation (eq. 1) or the Michaelis-Menten equation (eq. 2), which indicated that the reductase activity catalyzed by the AKR1C3 component in the #30 sample was clearly lower than that in the #10 sample (Fig. 3C). Taken together, these results demonstrated that the AKR1C3 C145Y carrier has decreased activity.

Discussion

AKR1C3 is highly expressed in the human liver, with an average of 14% of the total mRNA levels of members in the AKR and SDR superfamilies (Amai et al., 2020), and contributes to the detoxification or activation of various drugs (Barski et al., 2008; Ichida et al., 2023a). Interindividual variation in AKR1C3 activity would lead to variable efficacy or adverse reactions of drugs. Genetic polymorphism is one of the factors causing interindividual variations in drug response, but information on the effects of AKR1C3 genetic polymorphisms on enzyme activity is still scarce. We have previously shown that AKR1C3 highly contributes to the reduction reactions of acetohexamide, loxoprofen, and doxorubicin in HLC at 78%, 54%, and 51%, respectively (Ichida et al., 2023a). Therefore, this study aimed to clarify the effects of five types of SNPs on the AKR1C3 causing R66Q, E77G, C145Y, P180S, or R258C amino acid substitutions on the reductase activity for the above three representative substrates.

The CLint values of the reductase activities for all three substrates by recombinant R66Q were close to those of the wild type (Fig. 1, B–D; Table 3). This result is consistent with a previous report by Bains et al. (2010) evaluating the reductase activities of daunorubicin and doxorubicin. The CLint values of the reductase activities for all three substrates by recombinant E77G were also close to those of the wild type (Fig. 1, B–D; Table 3). Jakobsson et al. (2007) and Detlefsen et al. (2022) previously showed that recombinant E77G had similar or ∼20% lower activities than wild type for three steroid compounds, including exemestane, Δ4-androstenedione, and progesterone. Jakobsson et al. (2007) also reported that the heterozygotes of the E77G allele have 11% lower free serum testosterone levels than homozygotes of the wild type in Swedes. Thus, the SNP causing the amino acid substitution of E77G is likely to decrease the steroid metabolism, although its effects on drug metabolism are unlikely to be pronounced. The CLint values for acetohexamide, loxoprofen, and doxorubicin reductase activities of recombinant P180S were 30% higher, similar, and 20% lower than those of the wild type, respectively (Fig. 1, B–D; Table 3); thus, the effects of the P180S mutation on the activities appear to be dependent on substrates, although the effects are moderate. Bains et al. (2010) reported that the doxorubicin and daunorubicin reductase activities of recombinant P180S were 69% and 27% lower than those of wild type, respectively. The inconsistent effects of P180S on doxorubicin reductase activity might be caused by differences in the host cells for the expression systems between Bains et al. (2010) (Escherichia coli cells) and our study (Sf21 cells). The CLint values for acetohexamide and loxoprofen reductase activities by recombinant R258C were close to those of the wild type, and that for doxorubicin reductase activity by R258C was 50% higher than that of the wild type (Fig. 1, B–D; Table 3). Detlefsen et al. (2022) reported that Δ4-androstenedione reductase activity by recombinant R258C was close to that by wild type, but exemestane and progesterone reductase activities by recombinant R258C were 35% and 30% lower than those of wild type, respectively. Considering these results, arginine 258 may interact with a limited number of substrates to exert sufficient enzyme activity. Platt et al. (2016) reported that the CLint values of exemestane reductase activities by recombinant E77G, P180S, and R258C were extremely low, less than 5% of those of wild type, which was inconsistent with the reports by Detlefsen et al. (2022) and the present study because of no notable changes in CLint values of the activities by recombinant E77G, P180S, and R258C. It is unclear why such a large difference was observed between our results and those of Platt et al. (2016).

View this table:
TABLE 3

Kinetic parameters of acetohexamide, doxorubicin, and loxoprofen reduction by recombinant AKR1C3 variants

Recombinant C145Y showed 86%–94% lower CLint values for reductase activities for all three substrates than the wild type (Fig. 1, B–D; Table 3), indicating that the cysteine residue is critical for activity. Interestingly, the recombinant C145Y protein followed the Michaelis-Menten equation for the reduction of acetohexamide and loxoprofen, even though the wild-type protein and the other variants followed the substrate inhibition kinetics (Fig. 1, B and D). Considering the possibility that the cysteine at position 145 may form a disulfide bond, we examined the effect of 100 µM dithiothreitol, a reducing agent, on the reductase activity of the AKR1C3 wild type. Since acetohexamide reductase activity was decreased by only 12% in the presence of dithiothreitol (Supplemental Fig. 5), the disulfide bond through Cys 145 may not be relevant for the maintenance of AKR1C3 reductase activity. Adeniji et al. (2013) reported that the substrate binding pocket of AKR1C3 is composed of five regions, including the oxyanion site (Y55 and H117), the steroid channel (L54 and W227), and three subsites, SP1 (S118, N167, F306, F311, and Y319), SP2 (W86, S129, W227, and F311), and SP3 (Y24, E192, S217, S221, Q222, Y305, and F306). Based on the crystal structure of AKR1C3 (Supplemental Fig. 6), it was suggested that the cysteine residue at position 145 is located on the surface of the protein rather than in the substrate binding site. It is unclear how the cysteine residue at position 145 regulates acetohexamide and loxoprofen reductase activity, but the region around the cysteine residue at position 145 may act as an allosteric site to influence the structure of the substrate binding site.

The SNP g.55101G>A causing the amino acid substitution C145Y was observed in Koreans and Japanese individuals with allele frequencies of 6.8% and 3.6%, respectively, whereas it was not observed in Caucasians and African Americans (Table 4). The allele frequencies obtained in this study were consistent with those reported in the dbSNP database, which are 5.1% and 3.5% in Korean and Japanese individuals, respectively (https://www.ncbi.nlm.nih.gov/snp/rs28943579?horizontal_tab=true). The dbSNP database shows that this variant was not detected in South Asians; therefore, it seems that the variant is specifically detected in East Asians. In this study, the sequences of all exons and exon-intron junctions of the AKR1C3 gene were analyzed using genomic DNA from a carrier heterozygous for g.55101G>A (sample #24), but no additional SNPs were detected (data not shown); thus, there appears to be no linkage between g.55101G>A and other SNPs.

View this table:
TABLE 4

Allele frequency of g.55101G>A (rs28943579) in four populations

In individual Japanese liver samples, heterozygotes of g.55101G>A tended to have lower acetohexamide, loxoprofen, and doxorubicin reductase activities (Fig. 2, A–C). The reductase activities were measured at substrate concentrations of 20, 5, and 15 µM, respectively, which are the predicted hepatic maximum unbound drug concentrations (Cinlet,u,max) of these substrates as previously calculated by Ichida et al. (2023a) using the following calculation: Embedded Image where fu is the unbound fraction in the blood, Cmax is the maximum concentration in the blood, ka is the first-order rate constant for gastrointestinal absorption, fa is the fraction absorbed from the gastrointestinal tract into the portal vein, and QH is the hepatic blood flow rate. The predicted hepatic concentrations of loxoprofen and doxorubicin in the clinical situation are much lower than the Km values for wild type; therefore, it is conceivable that the AKR1C3 genotype could alter the serum concentration of these substrates. For acetohexamide, the predicted hepatic concentration was higher than the Km value of wild type but lower than the value of the C145Y variant. Thus, the serum concentration of the reduced metabolite, hydroxyhexamide, in homozygotes of g.55101G>A would be lower than that in homozygotes of wild type. The fractions metabolized to reduced products for acetohexamide, loxoprofen, and doxurubicin were 47%, 22%, and 22%, respectively (Galloway et al., 1967; Koo et al., 2005; Pérez-Blanco et al., 2016), and acetohexamide and loxoprofen are prodrugs activated by the reduction reaction; therefore, there may be a risk of drug toxicity or decreased efficacy in a g.55101G>A carrier. Furthermore, kinetic analysis of acetohexamide reductase activity using HLC samples showed that the CLint value of a heterozygote of g.55101G>A was lower than that of a noncarrier of g.55101G>A (Fig. 3; Table 5). Unfortunately, there were no homozygotes of g.55101G>A among the individual liver samples used in this study. The acetohexamide reductase activity at low concentrations, i.e. clinical concentrations, in homozygotes of g.55101G>A may obviously be lower than that in noncarriers or heterozygotes of g.55101G>A. The g.55101G>A allele would be a factor determining drug efficacy and toxicity. However, since AKR1C3 is an inducible enzyme, the decreased activity due to the mutation may be compensated in the situation where the AKR1C3 expression is greatly induced.

View this table:
TABLE 5

Kinetic parameters of acetohexamide reduction by No.10 and No.30 samples derived from a homozygote of AKR1C3 wild type and a heterozygote of g.55101G>A (rs28943579)

CLint value was calculated using an initial slope of the velocity versus the substrate concentration. The Km1, Vmax1, and Ki1 represent the component for AKR1C3, and the Km2 and Vmax2 represent the component for CBR1.

In conclusion, we found that the C145Y variant markedly decreased the enzyme activity of AKR1C3. In addition, HLC samples with g.55101G>A (AKR1C3 C145Y) tended to show low enzyme activity in the reduction reaction of clinical drugs. Since the g.55101G>A allele was specifically detected in East Asians, this allele may explain interindividual variability in drug efficacy and sensitivity to adverse effects related to AKR1C3 substrates in East Asians.

Data Availability

The authors declare that all the data supporting the findings of this study are available within the paper and its supplemental Material.

Authorship Contributions

Participated in research design: Takano, Fukami, Nakajima.

Conducted experiments: Takano.

Contributed new reagents or analytical tools: Takano, Ichida, Suzuki.

Performed data analysis: Takano, Fukami, Nakano, Nakajima.

Wrote or contributed to the writing of the manuscript: Takano, Fukami, Nakajima.

Footnotes

Abbreviations

AKR
aldo-keto reductase
CBR
carbonyl reductase
CLint
intrinsic clearance
dbSNP
Single Nucleotide Polymorphism Database
G6P
glucose-6-phosphate
HLC
human liver cytosol
Km
Michaelis-Menten constant
PCR
polymerase chain reaction
S
substrate concentration
SDR
short-chain dehydrogenase/reductase
SNP
single-nucleotide polymorphism
V
velocity
WT
wild type

References

  1. ↵
    1. Adeniji AO,
    2. Chen M, and
    3. Penning TM
    (2013) AKR1C3 as a target in castrate resistant prostate cancer. J Steroid Biochem Mol Biol 137:136–149.
    OpenUrlCrossRefPubMed
  2. ↵
    1. Adeniji AO,
    2. Twenter BM,
    3. Byrns MC,
    4. Jin Y,
    5. Chen M,
    6. Winkler JD, and
    7. Penning TM
    (2012) Development of potent and selective inhibitors of aldo-keto reductase 1C3 (type 5 17β-hydroxysteroid dehydrogenase) based on N-phenyl-aminobenzoates and their structure-activity relationships. J Med Chem 55:2311–2323.
    OpenUrlCrossRefPubMed
  3. ↵
    1. Amai K,
    2. Fukami T,
    3. Ichida H,
    4. Watanabe A,
    5. Nakano M,
    6. Watanabe K, and
    7. Nakajima M
    (2020) Quantitative analysis of mRNA expression levels of aldo-keto reductase and short-chain dehydrogenase/reductase isoforms in human livers. Drug Metab Pharmacokinet 35:539–547.
    OpenUrl
  4. ↵
    1. Bains OS,
    2. Grigliatti TA,
    3. Reid RE, and
    4. Riggs KW
    (2010) Naturally occurring variants of human aldo-keto reductases with reduced in vitro metabolism of daunorubicin and doxorubicin. J Pharmacol Exp Ther 335:533–545.
    OpenUrlAbstract/FREE Full Text
  5. ↵
    1. Barski OA,
    2. Tipparaju SM, and
    3. Bhatnagar A
    (2008) The aldo-keto reductase superfamily and its role in drug metabolism and detoxification. Drug Metab Rev 40:553–624.
    OpenUrlCrossRefPubMed
  6. ↵
    1. Bradford MM
    (1976) A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal Biochem 72:248–254.
    OpenUrlCrossRefPubMed
  7. ↵
    1. Byrns MC,
    2. Steckelbroeck S, and
    3. Penning TM
    (2008) An indomethacin analogue, N-(4-chlorobenzoyl)-melatonin, is a selective inhibitor of aldo-keto reductase 1C3 (type 2 3alpha-HSD, type 5 17beta-HSD, and prostaglandin F synthase), a potential target for the treatment of hormone dependent and hormone independent malignancies. Biochem Pharmacol 75:484–493.
    OpenUrlCrossRefPubMed
  8. ↵
    1. Detlefsen AJ,
    2. Wangtrakuldee P, and
    3. Penning TM
    (2022) Characterization of the major single nucleotide polymorphic variants of aldo-keto reductase 1C3 (type 5 17β-hydroxysteroid dehydrogenase). J Steroid Biochem Mol Biol 221:106121.
    OpenUrl
  9. ↵
    1. Evans WE and
    2. McLeod HL
    (2003) Pharmacogenomics--drug disposition, drug targets, and side effects. N Engl J Med 348:538–549.
    OpenUrlCrossRefPubMed
  10. ↵
    1. Fukami T,
    2. Takahashi S,
    3. Nakagawa N,
    4. Maruichi T,
    5. Nakajima M, and
    6. Yokoi T
    (2010) In vitro evaluation of inhibitory effects of antidiabetic and antihyperlipidemic drugs on human carboxylesterase activities. Drug Metab Dispos 38:2173–2178.
    OpenUrlAbstract/FREE Full Text
  11. ↵
    1. Fukami T,
    2. Yokoi T, and
    3. Nakajima M
    (2022) Non-P450 Drug-Metabolizing Enzymes: Contribution to Drug Disposition, Toxicity, and Development. Annu Rev Pharmacol Toxicol 62:405–425.
    OpenUrl
  12. ↵
    1. Galloway JA,
    2. McMahon RE,
    3. Culp HW,
    4. Marshall FJ, and
    5. Young EC
    (1967) Metabolism, blood levels and rate of excretion of acetohexamide in human subjects. Diabetes 16:118–127.
    OpenUrlAbstract/FREE Full Text
  13. ↵
    1. Hirosawa K,
    2. Fukami T,
    3. Tashiro K,
    4. Sakai Y,
    5. Kisui F,
    6. Nakano M, and
    7. Nakajima M
    (2021) Role of human arylacetamide deacetylase (AADAC) on hydrolysis of eslicarbazepine acetate and effects of AADAC genetic polymorphisms on hydrolase activity. Drug Metab Dispos 49:322–329.
    OpenUrlAbstract/FREE Full Text
  14. ↵
    1. Ichida H,
    2. Fukami T,
    3. Amai K,
    4. Suzuki K,
    5. Mishiro K,
    6. Takano S,
    7. Obuchi W,
    8. Zhang Z,
    9. Watanabe A,
    10. Nakano M, et al.
    (2023a) Quantitative evaluation of the contribution of each aldo-keto reductase and short-chain dehydrogenase/reductase isoform to reduction reactions of compounds containing a ketone group in the human liver. Drug Metab Dispos 51:17–28.
    OpenUrlAbstract/FREE Full Text
  15. ↵
    1. Ichida H,
    2. Fukami T,
    3. Kudo T,
    4. Mishiro K,
    5. Takano S,
    6. Nakano M,
    7. Morinaga G,
    8. Matsui A,
    9. Ishiguro N, and
    10. Nakajima M
    (2023b) Identification of HSD17B12 as an enzyme catalyzing drug reduction reactions through investigation of nabumetone metabolism. Arch Biochem Biophys 736:109536.
    OpenUrl
  16. ↵
    1. Jakobsson J,
    2. Palonek E,
    3. Lorentzon M,
    4. Ohlsson C,
    5. Rane A, and
    6. Ekström L
    (2007) A novel polymorphism in the 17β-hydroxysteroid dehydrogenase type 5 (aldo-keto reductase 1C3) gene is associated with lower serum testosterone levels in caucasian men. Pharmacogenomics J 7:282–289.
    OpenUrlCrossRefPubMed
  17. ↵
    1. Jung KA,
    2. Choi BH,
    3. Nam CW,
    4. Song M,
    5. Kim ST,
    6. Lee JY, and
    7. Kwak MK
    (2013) Identification of aldo-keto reductases as NRF2-target marker genes in human cells. Toxicol Lett 218:39–49.
    OpenUrlCrossRefPubMed
  18. ↵
    1. Kassner N,
    2. Huse K,
    3. Martin HJ,
    4. Gödtel-Armbrust U,
    5. Metzger A,
    6. Meineke I,
    7. Brockmöller J,
    8. Klein K,
    9. Zanger UM,
    10. Maser E, et al.
    (2008) Carbonyl reductase 1 is a predominant doxorubicin reductase in the human liver. Drug Metab Dispos 36:2113–2120.
    OpenUrlAbstract/FREE Full Text
  19. ↵
    1. Kobayashi Y,
    2. Fukami T,
    3. Nakajima A,
    4. Watanabe A,
    5. Nakajima M, and
    6. Yokoi T
    (2012) Species differences in tissue distribution and enzyme activities of arylacetamide deacetylase in human, rat, and mouse. Drug Metab Dispos 40:671–679.
    OpenUrlAbstract/FREE Full Text
  20. ↵
    1. Koo TS,
    2. Kim DH,
    3. Ahn SH,
    4. Kim KP,
    5. Kim IW,
    6. Seo SY,
    7. Suh YG,
    8. Kim DD,
    9. Shim CK, and
    10. Chung SJ
    (2005) Comparison of pharmacokinetics of loxoprofen and its active metabolites after an intravenous, intramuscular, and oral administration of loxoprofen in rats: evidence for extrahepatic metabolism. J Pharm Sci 94:2187–2197.
    OpenUrlPubMed
  21. ↵
    1. Laemmli UK
    (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227:680–685.
    OpenUrlCrossRefPubMed
  22. ↵
    1. Malátková P,
    2. Maser E, and
    3. Wsól V
    (2010) Human carbonyl reductases. Curr Drug Metab 11:639–658.
    OpenUrlCrossRefPubMed
  23. ↵
    1. Malátková P,
    2. Sokolová S,
    3. Chocholoušová Havlíková L, and
    4. Wsól V
    (2016) Carbonyl reduction of warfarin: Identification and characterization of human warfarin reductases. Biochem Pharmacol 109:83–90.
    OpenUrl
  24. ↵
    1. Matsunaga T,
    2. Shintani S, and
    3. Hara A
    (2006) Multiplicity of mammalian reductases for xenobiotic carbonyl compounds. Drug Metab Pharmacokinet 21:1–18.
    OpenUrlCrossRefPubMed
  25. ↵
    1. Ohara H,
    2. Miyabe Y,
    3. Deyashiki Y,
    4. Matsuura K, and
    5. Hara A
    (1995) Reduction of drug ketones by dihydrodiol dehydrogenases, carbonyl reductase and aldehyde reductase of human liver. Biochem Pharmacol 50:221–227.
    OpenUrlCrossRefPubMed
  26. ↵
    1. Penning TM
    (2015) The aldo-keto reductases (AKRs): Overview. Chem Biol Interact 234:236–246.
    OpenUrlCrossRefPubMed
  27. ↵
    1. Pérez-Blanco JS,
    2. Santos-Buelga D,
    3. Fernández de Gatta MD,
    4. Hernández-Rivas JM,
    5. Martín A, and
    6. García MJ
    (2016) Population pharmacokinetics of doxorubicin and doxorubicinol in patients diagnosed with non-Hodgkin’s lymphoma. Br J Clin Pharmacol 82:1517–1527.
    OpenUrl
  28. ↵
    1. Persson B and
    2. Kallberg Y
    (2013) Classification and nomenclature of the superfamily of short-chain dehydrogenases/reductases (SDRs). Chem Biol Interact 202:111–115.
    OpenUrlCrossRefPubMed
  29. ↵
    1. Platt A,
    2. Xia Z,
    3. Liu Y,
    4. Chen G, and
    5. Lazarus P
    (2016) Impact of nonsynonymous single nucleotide polymorphisms on in-vitro metabolism of exemestane by hepatic cytosolic reductases. Pharmacogenet Genomics 26:370–380.
    OpenUrl
  30. ↵
    1. Taavitsainen P,
    2. Prien O,
    3. Kähkönen M,
    4. Niehues M,
    5. Korjamo T,
    6. Denner K,
    7. Nykänen P,
    8. Vuorela A,
    9. Jungmann NA,
    10. von Bühler CJ, et al.
    (2021) Metabolism and Mass Balance of the Novel Nonsteroidal Androgen Receptor Inhibitor Darolutamide in Humans. Drug Metab Dispos 49:420–433.
    OpenUrlAbstract/FREE Full Text
  31. ↵
    1. Yamashita N,
    2. Kanno Y,
    3. Saito N,
    4. Terai K,
    5. Sanada N,
    6. Kizu R,
    7. Hiruta N,
    8. Park Y,
    9. Bujo H, and
    10. Nemoto K
    (2019) Aryl hydrocarbon receptor counteracts pharmacological efficacy of doxorubicin via enhanced AKR1C3 expression in triple negative breast cancer cells. Biochem Biophys Res Commun 516:693–698.
    OpenUrl
    1. Zhou Q,
    2. Tian W,
    3. Jiang Z,
    4. Huang T,
    5. Ge C,
    6. Liu T,
    7. Zhao F,
    8. Chen T,
    9. Cui Y,
    10. Li H, et al.
    (2021) A positive feedback loop of AKR1C3-mediated activation of NF-κB and STAT3 facilitates proliferation and metastasis in hepatocellular carcinoma. Cancer Res 81:1361–1374.
    OpenUrlAbstract/FREE Full Text
Back to top

In this issue

Download PDF
Article Alerts
Email Article
Citation Tools

Research ArticleArticle

Effects of AKR1C3 Variants on Reductase Activity

Shiori Takano, Tatsuki Fukami, Hiroyuki Ichida, Kohei Suzuki, Masataka Nakano and Miki Nakajima
Drug Metabolism and Disposition September 1, 2023, 51 (9) 1188-1195; DOI: https://doi.org/10.1124/dmd.123.001264
Twitter logo Facebook logo Mendeley logo

Jump to section

Advertisement